My thoughts on "PID Controllers - Part 3A of Why I don't recommend one."
Topic 3: What is a PID controller?
--3A: Proportional control--
Note: I'm specifically saying I don't recommend a PID controller, not that I don't want to control the temperature! Also, there is nothing wrong with using a PID controller; I'm just saying it's not needed!
I'm going to discuss the following
1 Why are we interested in controlling smoke-chamber temperature?
2 What are some different methods in controlling temperature?
3 What is a PID controller?
3A - Proportional
3B - Integral
3C - Derivative
4 Why would someone use a PID controller?
5 Is a PID controller needed?
6 What is the optimum controller for controlling a pellet smoker?
7 Why I don't recommend a PID controller
Previous: https://www.smokingmeatforums.com/threads/pid-controllers-part-2-of-why-i-dont-recomend-one.275118/
I will probably break this portion up into three (3) separate entries: Proportional, Integral, Derivative (PID).
In order to present this clearly, I'm using sources to assist me rather than just based on memory ;-) I'll provide a 'reference' list at the bottom.
---Background---
A controller tries to control a process against an error measured against the desired value and the observed value. A simple timer controller, such as that provided by the P "smoke" setting isn't a true controller, as it does not measure any error against a desired temperature setpoint. It simply turns on and off based on a time value, while completely ignoring any resulting from these on - off actions.
The PID controller has an origin based on a centrifugal governor (used rotating weights to control. I have seen them in old movies, but always thought they were just props, not real!). Research indicated one of the first usages of a controller was to control the gap between millstones in windmills! This was a proportional controller.
---Proportional Control---
Varying loads would increase the error observed vs the setpoint on a proportional controller.
The proportional control produces an output that is proportional to the current error, where error is the difference between the observed and desired output. In the case of a proportional temperature controller, the error is the difference between the observed temperature and the temperature setpoint. The output could be used to control (as examples): A) fuel supply (e.g., an auger): rpm, or time on ratio to time off; B)air damper: open more when below setpoint, close more when above setpoint; C) electric heater: more current when below setpoint, less current when above setpoint; D) a fan aiding combustion of wood pellets: fan on when burning, fan off when pellets consumed; E) a fan: increase rpm for more heat, less rpm for less heat.
Most proportional controllers have some method for varying the output response to the input error; this is the amount of 'gain' applied to the output. The higher the gain, the more the difference a given amount of error will produce. If the gain is too high, the system can become unstable - overshooting and undershooting it's setpoint by a large margin. This appears indicative of a problem with the controller used in the Pit Boss 820D, in my opnion - I'm not aware of any method available to me to change the gain setting on this controller.
So, a proportional controller will generate an output that should directly correlate to the error. The error in this case is the difference between the desired temperature (setpoint) and the observed temperature. If only an on/off control is available, then the control will be the length of time on vs length of time off, or simply on when more heat is needed and off when no heat is needed - the household thermostat.
Previous: https://www.smokingmeatforums.com/threads/pid-controllers-part-2-of-why-i-dont-recomend-one.275118/
Next: Part 3B: Why would someone use a PID controller? - will follow soon.
My background: 10 years in process control at nuclear power plants, as a technician and engineer, and 41 years in software. Smoking BBQ since 1989.
Banjo
--References--
https://en.wikipedia.org/wiki/PID_controller
Topic 3: What is a PID controller?
--3A: Proportional control--
Note: I'm specifically saying I don't recommend a PID controller, not that I don't want to control the temperature! Also, there is nothing wrong with using a PID controller; I'm just saying it's not needed!
I'm going to discuss the following
1 Why are we interested in controlling smoke-chamber temperature?
2 What are some different methods in controlling temperature?
3 What is a PID controller?
3A - Proportional
3B - Integral
3C - Derivative
4 Why would someone use a PID controller?
5 Is a PID controller needed?
6 What is the optimum controller for controlling a pellet smoker?
7 Why I don't recommend a PID controller
Previous: https://www.smokingmeatforums.com/threads/pid-controllers-part-2-of-why-i-dont-recomend-one.275118/
I will probably break this portion up into three (3) separate entries: Proportional, Integral, Derivative (PID).
In order to present this clearly, I'm using sources to assist me rather than just based on memory ;-) I'll provide a 'reference' list at the bottom.
---Background---
A controller tries to control a process against an error measured against the desired value and the observed value. A simple timer controller, such as that provided by the P "smoke" setting isn't a true controller, as it does not measure any error against a desired temperature setpoint. It simply turns on and off based on a time value, while completely ignoring any resulting from these on - off actions.
The PID controller has an origin based on a centrifugal governor (used rotating weights to control. I have seen them in old movies, but always thought they were just props, not real!). Research indicated one of the first usages of a controller was to control the gap between millstones in windmills! This was a proportional controller.
---Proportional Control---
Varying loads would increase the error observed vs the setpoint on a proportional controller.
The proportional control produces an output that is proportional to the current error, where error is the difference between the observed and desired output. In the case of a proportional temperature controller, the error is the difference between the observed temperature and the temperature setpoint. The output could be used to control (as examples): A) fuel supply (e.g., an auger): rpm, or time on ratio to time off; B)air damper: open more when below setpoint, close more when above setpoint; C) electric heater: more current when below setpoint, less current when above setpoint; D) a fan aiding combustion of wood pellets: fan on when burning, fan off when pellets consumed; E) a fan: increase rpm for more heat, less rpm for less heat.
Most proportional controllers have some method for varying the output response to the input error; this is the amount of 'gain' applied to the output. The higher the gain, the more the difference a given amount of error will produce. If the gain is too high, the system can become unstable - overshooting and undershooting it's setpoint by a large margin. This appears indicative of a problem with the controller used in the Pit Boss 820D, in my opnion - I'm not aware of any method available to me to change the gain setting on this controller.
So, a proportional controller will generate an output that should directly correlate to the error. The error in this case is the difference between the desired temperature (setpoint) and the observed temperature. If only an on/off control is available, then the control will be the length of time on vs length of time off, or simply on when more heat is needed and off when no heat is needed - the household thermostat.
Previous: https://www.smokingmeatforums.com/threads/pid-controllers-part-2-of-why-i-dont-recomend-one.275118/
Next: Part 3B: Why would someone use a PID controller? - will follow soon.
My background: 10 years in process control at nuclear power plants, as a technician and engineer, and 41 years in software. Smoking BBQ since 1989.
Banjo
--References--
https://en.wikipedia.org/wiki/PID_controller
Last edited:
